File preview
The TU Dortmund Cyber-Physical Systems Program A Step Towards Multi-Disciplinary Education
Peter Marwedel, Wolfgang Rhode, Katharina Morik
TU Dortmund 44221 Dortmund, Germany http://ls12-www.cs.tu-dortmund.de
Abstract—In this paper we propose an integrated curriculum for a cyber-physical system (CPS) program. The various possible focus areas of CPS are included via electives. The program includes topics from computer science, physics, electrical engineering, mathematics, and —for some electives— also mechanical and control-engineering. It is essential that this program includes undergraduate as well as graduate level education. The program provides an attractive, interdisciplinary education in a field of rapidly growing importance. In this way, we describe an example of a curriculum which in the future should contribute to better (more sound) CPS engineering.
first feed-back. The paper will conclude with a summary in section V. II. P ROPOSAL FOR A CPS CURRICULUM A. Principles It has been observed that is essential that the fundamental principles of the involved science and engineering disciplines are included in curricula. Fashionable new topics come and go. Therefore, we consider it necessary to include the fundamental principles of relevant sciences and to refrain from loading the new program with topics which might not be essential in the long run. Involved sciences and engineering disciplines include computer science, physics, electrical engineering, mathematics, and possibly also mechanical and control engineering. According to the European Bologna standards [3], European educational programs have to be partitioned into an undergraduate and a graduate phase and have to use a credit point system. The minimum duration of the undergraduate phase is three years. Regulations require a workload of 900 hours per term and a workload of 30 hours per credit point (CP). The partitioning into the two phases may vary, but a partitioning into 6 + 4 terms is currently the rule, at least in Germany. One internship term is sometimes optional, especially in engineering disciplines. We believe that it is important to start the CPS education already at the undergraduate level, since an education starting at the graduate level would not provide enough headroom for the essential topics and would leave the students in their original discipline, i.e. a CS undergraduate would not become a real CPS student during the graduate phase. In our proposal, we stick to the 6 + 4 partitioning. Due to the limited number of credits that are available for each discipline, we decided to include foundations of the disciplines in the undergraduate program, whereas some advanced formal topics are allocated to the graduate program. In this way, basic skills are available for students who enter the job market after receiving their Bachelor’s degree. In the following, we will describe the topics which we consider essential. The description will be based on courses even though other universities might decide to repartition the topics into courses in a different way. B. Undergraduate level topics 1) Computer Science (CS): All cyber-physical systems that we can think of involve some amount of software. We believe
I. I NTRODUCTION In recent years, information and communication technologies (ICT) have increasingly been integrated into products and systems. In this way, the application of ICT to office automation has been extended much beyond its previously dominating area. ICT components embedded into other products and systems have been called embedded systems: “Embedded systems are information processing systems embedded into a larger product” [1]. Even more recently, the trend towards a tight integration between ICT and the physical environment has been strengthened. We can observe an increased emphasis on modeling the physical environment together with the ICT components. This has led to the introduction of the term “cyber-physical systems” (CPS). “Cyber-Physical Systems (CPS) are integrations of computation with physical processes” [2]. Previous discussions at the Workshop on Embedded System Education (WESE) have come to the conclusion that teaching embedded system design is difficult: Embedded system education is typically implemented as a concentration (specialization) within some other well-known program such as computer science or electrical engineering. In these cases, the advantages of using well-known degree names is given preference over more headroom for courses in all relevant areas and more specialized courses. With the introduction of CPS, we are passing the limits of this approach. It is therefore necessary to consider the introduction of an integrated CPS program. The attractiveness of CPS as a term also allows us to introduce a separate program. Currently, we are designing the new program, scheduled to start about one year from now. The remainder of this paper is structured as follows: in section II, we will be describing our approach for designing a CPS curriculum. A brief look at related work will be the topic of section III. Section IV will contain a short view on
that the capability to design software remains essential in the years to come. According to our experience with classical computer science programs, it is necessary to include a number of software-related courses. Also, students need a good understanding of technological platforms on which software can be executed. The topics and corresponding courses are listed in fig. I.
Name Algorithms & programming Computer organization Operating systems Computer networks Embedded systems Information systems Introduction to data analysis Content data structures, complexity, exemplary programming Boolean logic, computer architecture & arithmetic processes, threads, resource management protocols, standards as defined in [1] data bases classification, regression, clustering, frequent set mining, analysis of streaming sensor data Credits 24 9 5 5 9 4 9
concentration. For our point of view, students opting for some other concentration do not need more knowledge in EE than what is listed in Table III. 4) Mathematics: For the education in mathematics, we are in a dilemma: CS education requires advanced knowledge in discrete mathematics while physics mostly requires classical engineering mathematics. We believe that engineering mathematics is essential for physics and resolve the conflict in favor of physics.
Name Engineering math 1 Engineering math 2 Engineering math 3 Content series, linear equations integrals, derivatives vector analysis, transforms, Eigen value problems TABLE IV M ATH TOPICS Credits 9 9 9
TABLE I C OMPUTER SCIENCE TOPICS
We retain teaching programming and algorithms in some depth. Compared to a standard CS curriculum, we have to drop teaching a second programming paradigm (such as logic programming) and must shift some more theoretical content (such as formal automata theory) to the graduate level. Computer organization is considered being essential for CPS education. 2) Physics: It is obvious from the name CPS, that physics must be included in a CPS program.
Name Experimental physics 1 Experimental physics 2 Experimental physics 3 Content masses, engineering mechanics, relativity electrodynamics quantum & particle physics TABLE II P HYSICS TOPICS Credits 9 9 9
5) CPS: All the above courses are courses which do already exist and for which students in other programs can enroll themselves as well. In this way, we take into account that budget constraints will not allow us to hire new faculty members for CPS. Nevertheless, the integration of the different disciplines must be demonstrated to the students. Therefore, some integrated courses are required. We envision two contexts in which such an integration is feasible, as shown in table V. In addition the software project may focus on CPS.
Name Integrated CPS course Bachelor Thesis Content Linking ≥ 2 disciplines CPS-related topic Credits 11 15
TABLE V I NTEGRATED CPS TOPICS
In this area, we have included classical experimental physics topics. Compared to a standard physics curriculum, we must shift some more theoretical content to the graduate level. We assume that this also helps to maintain a high level of motivation among undergraduates. 3) Electrical Engineering: Electrical engineering (EE) provides indispensable techniques for implementing CPS. Therefore, EE topics are also required (see table III).
Name Fundamentals of EE Electronics Content electrical networks electronic circuits1 TABLE III EE TOPICS Credits 9 9
6) Concentrations: It is one of the characteristics of CPS that different communities are using the term in different ways, each of which is attractive by itself. Instead of selecting one of these ways, we offer the different interpretations as electives grouped into concentrations (focus areas). Students must obtain 18 credit points from electives. We recommend to stay within one focus area, but mixing and matching is also feasible. Possible concentrations are described next. a) Embedded Systems: Embedded systems (ES) in the sense defined in the introduction remain important, despite the more comprehensive context of CPS. Students may chose between different undergraduate courses offered by the department of electrical engineering. (see Table VI).
Name Control Communication Mechatronics Content Control circuits, stability coding, signal to noise ratio hybrid systems Credits 9 9 9
TABLE VI E MBEDDED SYSTEMS TOPICS
The total number of credit points for EE may be less than expected for an embedded systems program. This is motivated by the fact that embedded systems are included as a possible
b) Big data: Big data (how to handle large amounts of data) is one of the hot topics. Data resulting from physical
experiments is one special case. We include this as a concentration area, even though this area is typically not mentioned as a possible interpretation of CPS. However, techniques for analyzing massive data from such experiments is certainly a case of integrating cyber and physics.
Name Distributed data mining TBD Content data Analysis with Hadoop, web & (social) network mining .. TABLE VII B IG DATA TOPICS Credits 9 ..
Mathematics 15% Electrical engineering 5% CPS course 6.1% Thesis 8.3% Electives 10% Computer science 36.1%
c) Internet of things: The integration of the internet and the world of embedded systems is also one of the hottest topics these days. In this area, linking sensors to the internet is considered important. The area is emphasized by some German initiatives [4]. At Dortmund, special importance of this area results from the fact that TU Dortmund has logistics as a focus area and modern logistics is very much placing an emphasis on the internet of things. Introduction to logistics as well as other courses offered by the mechanical engineering department are on the list of electives. d) Robotics and control: Robotics and control for cyberphysical systems is a topic covered in many of the CPS research programs, such as NSF’s CPS program [5]. We include this area as a concentration. e) Systems of systems: In this area, we are considering the linking of of objects which by themselves are already systems. For example, we consider the linking of power stations to form a whole network of power stations and power consumers. Also, we consider networks of cars, trains or planes. Network technology is an essential ingredient of this area. Also, we must design means for locally safe operation despite possibly failing networks. At TU Dortmund, power systems would be a possible focus. Systems of systems are one of the focus areas of the European Commission. f) Life sciences: Our education in life sciences as a concentration focuses on the analysis of large amounts of medical data. Data analysis techniques form the kernel of this specialization. There would be some overlap with the specialization on big data, but there would also be some specific topics, on some cases linked to genetics and related topics in biology. 7) Global view: We plan to start with a list of electives taken from three focus areas and add more when the program becomes more mature. Note that all of the areas are based on some methodology which can be applied in various application domains. We could also image to have concentrations for certain application domains, like • transportation (automotive, avionics, rail), • power generation and distribution, • logistics, disaster planning and recovery, • automation, smart buildings, • bio-technology, health industries. Fig. 1 demonstrates the relative percentages of the involved disciplines.
Physics 19.4%
Fig. 1. Global percentages for undergraduate workload
C. Graduate courses There are many theoretical aspects of CPS. Our curriculum design is based on the idea of covering such aspects during the graduate education, both for computer science and physics. More depth and supplementary material is provided in CPSrelated topics. The material the theory of automata and parsing, as shown in Table VIII.
Name Topics in theoretical CS Elective: formal methods Elective: modeling & simulation Content computational game theory, complexity, automata, languages logic, formal methods for system design, functional programming discrete event systems, queuing Credits 8 9 9
TABLE VIII S UPPLEMENTARY CS MATERIAL DURING THE GRADUATE PHASE
We consider the emphasis on the analysis of physical experiments as one of our characteristics. hence, more depth is also included for physics (see Table IX).
Name Structure of matter Theoretical physics 1 Theoretical physics 2 Content solid state physics classical physics quantum mechanics Credits 9 9 9
TABLE IX S UPPLEMENTARY PHYSICS MATERIAL DURING THE GRADUATE PHASE
In addition, 24 credit points are reserved for electives. For example, advanced courses on data analysis, embedded system design or control may be selected. 30 credit points are allocated to the Master thesis, 6 credit points have to be earned in a project and 7 credit points are reserved for non-technical skills. A global break-down of the graduate phase is shown in Fig. 2 Note that many of the more theoretical courses are comprised in the graduate education. In this way, students are prepared to go for their PhD.
Non-technical skills 5.8% CPS project 5% Thesis 25% Physics supplements 22.5% Electives 20% CS supplements 21.6%
•
•
Fig. 2. Global percentages for graduate workload
III. R ELATED WORK Much of the experience with CPS education is based on embedded systems education. The workshops on embedded system education provide an overview over work in this area (see [6] for a recent edition). New educational CPS programs are being designed at several universities. At the time of writing, no complete overview of such programs is available. We expect a more complete view of the current situation to be one of the key results of the first workshop on CPS education. The book by Edward Lee [7] focuses on CPS education. IV. E VALUATION AND FEED - BACK Due to the fact that CPS is a brand-new area, no feed-back or evaluations by students exist for this area. However, the following evaluation and feed-back exists: • The call of the first workshop on CPS education describes a list of outcomes of a good CPS curriculum [8]. Due to the multidisciplinary nature of the proposed curriculum, we have addressed in particular the following suggested outcomes: A. “An ability to apply mathematical models of physical systems, cyber systems, and their composition”. D. “An ability to function effectively on multidisciplinary teams spanning cyber and physical domains.” E. “An ability to identify, formulate, and solve engineering problems that have both cyber and physical aspects.” G. “An ability to communicate effectively across cyber and physical domains.” J. “A knowledge of contemporary issues with cyber-physical systems.” K. “An ability to select and use appropriate techniques, skills, and modern engineering tools that span the cyber and physical domains.” Many of the other outcomes are also addressed, at least to some extent.
Our ES courses are designed to provide embedded systems foundations for cyber-physical systems. Therefore, experiences gained in this area can be applied. ES courses, even though they are electives, are very popular among students. Typically more than 100 students attend the German course and a smaller number the corresponding English course. No major difficulties have been reported. It is, however, essential that a course on computer organization is included in our program. Such a course is not present even in plans for a new curriculum for computer science education [9]. Also, most students attend a hardware lab. Prerequisites which might be lacking among CS students are included in the appendix of our ES text book [1]. In meetings where this program was presented, there was an unanimous agreement among colleagues to introduce this program. Colleagues wanted to introduce this program as soon a possible.
V. C ONCLUSION In this paper, we have presented our scheme for an integrated CPS curriculum. From the list of presented topics, it is clear that such an integrated program is required in order to allocate a sufficient workload. Such a program has to start already at the undergraduate level. With such an integrated program, there is headroom for teaching the fundamentals of the involved areas. However, even with such an integrated program it is not possible to include all the flavors of CPS. Instead of restricting ourselves to a particular flavor, we include these as areas for electives. ACKNOWLEDGMENT Part of the work on this paper has been supported by Deutsche Forschungsgemeinschaft (DFG) within the Collaborative Research Center SFB 876 “Providing Information by Resource-Constrained Analysis”. R EFERENCES
[1] P. Marwedel, Embedded System Design – Embedded Systems Foundations of Cyber-Physical Systems. Springer, 2011. [2] E. A. Lee, “Cyber physical systems: Design challenges,” in International Symposium on Object/Component/Service-Oriented RealTime Distributed Computing (ISORC), May 2008, invited Paper. [Online]. Available: http://chess.eecs.berkeley.edu/pubs/427.html [3] European Ministers of Education, “The bologna declaration of 19 june 1999,” http:// www.bologna-bergen2005.no/ Docs/ 00-Main doc/ 990719BOLOGNA DECLARATION.PDF, 1999. [4] acatech, “Cyber-physical systems — driving force for innovation in mobility, health, energy and production,” http:// www.fortiss.org/ fileadmin/ uploads/ projects/ agendaCPS Position-paper.pdf , 2011. [5] CPS Virtual Organization, “home page,” http:// www.cps-vo.org. [6] P. Marwedel, J. Jackson, and K. Ricks, Eds., WESE ’11: Proceedings of the 6th Workshop on Embedded Systems Education. New York, NY, USA: ACM, 2011. [7] E. A. Lee and S. A. Seshia, “Introduction to embedded systems, a cyber-physical systems approach,” http:// LeeSeshia.org, ISBN 978-0-55770857-4, 2011. [8] “First workshop on cyber-physical systems education — call for papers,” http://cps-vo.org/group/edu/workshop. [9] M. Sahami, D. Grossman, R. LeBlanc, and R. Seker, “Computer science curriculum 2013: Curricular guidelines for the next decade,” http:// cra.org/ uploads/ documents/ resources/ snowbird2012 slides/ sahami.pdf , 2013.
Peter Marwedel, Wolfgang Rhode, Katharina Morik
TU Dortmund 44221 Dortmund, Germany http://ls12-www.cs.tu-dortmund.de
Abstract—In this paper we propose an integrated curriculum for a cyber-physical system (CPS) program. The various possible focus areas of CPS are included via electives. The program includes topics from computer science, physics, electrical engineering, mathematics, and —for some electives— also mechanical and control-engineering. It is essential that this program includes undergraduate as well as graduate level education. The program provides an attractive, interdisciplinary education in a field of rapidly growing importance. In this way, we describe an example of a curriculum which in the future should contribute to better (more sound) CPS engineering.
first feed-back. The paper will conclude with a summary in section V. II. P ROPOSAL FOR A CPS CURRICULUM A. Principles It has been observed that is essential that the fundamental principles of the involved science and engineering disciplines are included in curricula. Fashionable new topics come and go. Therefore, we consider it necessary to include the fundamental principles of relevant sciences and to refrain from loading the new program with topics which might not be essential in the long run. Involved sciences and engineering disciplines include computer science, physics, electrical engineering, mathematics, and possibly also mechanical and control engineering. According to the European Bologna standards [3], European educational programs have to be partitioned into an undergraduate and a graduate phase and have to use a credit point system. The minimum duration of the undergraduate phase is three years. Regulations require a workload of 900 hours per term and a workload of 30 hours per credit point (CP). The partitioning into the two phases may vary, but a partitioning into 6 + 4 terms is currently the rule, at least in Germany. One internship term is sometimes optional, especially in engineering disciplines. We believe that it is important to start the CPS education already at the undergraduate level, since an education starting at the graduate level would not provide enough headroom for the essential topics and would leave the students in their original discipline, i.e. a CS undergraduate would not become a real CPS student during the graduate phase. In our proposal, we stick to the 6 + 4 partitioning. Due to the limited number of credits that are available for each discipline, we decided to include foundations of the disciplines in the undergraduate program, whereas some advanced formal topics are allocated to the graduate program. In this way, basic skills are available for students who enter the job market after receiving their Bachelor’s degree. In the following, we will describe the topics which we consider essential. The description will be based on courses even though other universities might decide to repartition the topics into courses in a different way. B. Undergraduate level topics 1) Computer Science (CS): All cyber-physical systems that we can think of involve some amount of software. We believe
I. I NTRODUCTION In recent years, information and communication technologies (ICT) have increasingly been integrated into products and systems. In this way, the application of ICT to office automation has been extended much beyond its previously dominating area. ICT components embedded into other products and systems have been called embedded systems: “Embedded systems are information processing systems embedded into a larger product” [1]. Even more recently, the trend towards a tight integration between ICT and the physical environment has been strengthened. We can observe an increased emphasis on modeling the physical environment together with the ICT components. This has led to the introduction of the term “cyber-physical systems” (CPS). “Cyber-Physical Systems (CPS) are integrations of computation with physical processes” [2]. Previous discussions at the Workshop on Embedded System Education (WESE) have come to the conclusion that teaching embedded system design is difficult: Embedded system education is typically implemented as a concentration (specialization) within some other well-known program such as computer science or electrical engineering. In these cases, the advantages of using well-known degree names is given preference over more headroom for courses in all relevant areas and more specialized courses. With the introduction of CPS, we are passing the limits of this approach. It is therefore necessary to consider the introduction of an integrated CPS program. The attractiveness of CPS as a term also allows us to introduce a separate program. Currently, we are designing the new program, scheduled to start about one year from now. The remainder of this paper is structured as follows: in section II, we will be describing our approach for designing a CPS curriculum. A brief look at related work will be the topic of section III. Section IV will contain a short view on
that the capability to design software remains essential in the years to come. According to our experience with classical computer science programs, it is necessary to include a number of software-related courses. Also, students need a good understanding of technological platforms on which software can be executed. The topics and corresponding courses are listed in fig. I.
Name Algorithms & programming Computer organization Operating systems Computer networks Embedded systems Information systems Introduction to data analysis Content data structures, complexity, exemplary programming Boolean logic, computer architecture & arithmetic processes, threads, resource management protocols, standards as defined in [1] data bases classification, regression, clustering, frequent set mining, analysis of streaming sensor data Credits 24 9 5 5 9 4 9
concentration. For our point of view, students opting for some other concentration do not need more knowledge in EE than what is listed in Table III. 4) Mathematics: For the education in mathematics, we are in a dilemma: CS education requires advanced knowledge in discrete mathematics while physics mostly requires classical engineering mathematics. We believe that engineering mathematics is essential for physics and resolve the conflict in favor of physics.
Name Engineering math 1 Engineering math 2 Engineering math 3 Content series, linear equations integrals, derivatives vector analysis, transforms, Eigen value problems TABLE IV M ATH TOPICS Credits 9 9 9
TABLE I C OMPUTER SCIENCE TOPICS
We retain teaching programming and algorithms in some depth. Compared to a standard CS curriculum, we have to drop teaching a second programming paradigm (such as logic programming) and must shift some more theoretical content (such as formal automata theory) to the graduate level. Computer organization is considered being essential for CPS education. 2) Physics: It is obvious from the name CPS, that physics must be included in a CPS program.
Name Experimental physics 1 Experimental physics 2 Experimental physics 3 Content masses, engineering mechanics, relativity electrodynamics quantum & particle physics TABLE II P HYSICS TOPICS Credits 9 9 9
5) CPS: All the above courses are courses which do already exist and for which students in other programs can enroll themselves as well. In this way, we take into account that budget constraints will not allow us to hire new faculty members for CPS. Nevertheless, the integration of the different disciplines must be demonstrated to the students. Therefore, some integrated courses are required. We envision two contexts in which such an integration is feasible, as shown in table V. In addition the software project may focus on CPS.
Name Integrated CPS course Bachelor Thesis Content Linking ≥ 2 disciplines CPS-related topic Credits 11 15
TABLE V I NTEGRATED CPS TOPICS
In this area, we have included classical experimental physics topics. Compared to a standard physics curriculum, we must shift some more theoretical content to the graduate level. We assume that this also helps to maintain a high level of motivation among undergraduates. 3) Electrical Engineering: Electrical engineering (EE) provides indispensable techniques for implementing CPS. Therefore, EE topics are also required (see table III).
Name Fundamentals of EE Electronics Content electrical networks electronic circuits1 TABLE III EE TOPICS Credits 9 9
6) Concentrations: It is one of the characteristics of CPS that different communities are using the term in different ways, each of which is attractive by itself. Instead of selecting one of these ways, we offer the different interpretations as electives grouped into concentrations (focus areas). Students must obtain 18 credit points from electives. We recommend to stay within one focus area, but mixing and matching is also feasible. Possible concentrations are described next. a) Embedded Systems: Embedded systems (ES) in the sense defined in the introduction remain important, despite the more comprehensive context of CPS. Students may chose between different undergraduate courses offered by the department of electrical engineering. (see Table VI).
Name Control Communication Mechatronics Content Control circuits, stability coding, signal to noise ratio hybrid systems Credits 9 9 9
TABLE VI E MBEDDED SYSTEMS TOPICS
The total number of credit points for EE may be less than expected for an embedded systems program. This is motivated by the fact that embedded systems are included as a possible
b) Big data: Big data (how to handle large amounts of data) is one of the hot topics. Data resulting from physical
experiments is one special case. We include this as a concentration area, even though this area is typically not mentioned as a possible interpretation of CPS. However, techniques for analyzing massive data from such experiments is certainly a case of integrating cyber and physics.
Name Distributed data mining TBD Content data Analysis with Hadoop, web & (social) network mining .. TABLE VII B IG DATA TOPICS Credits 9 ..
Mathematics 15% Electrical engineering 5% CPS course 6.1% Thesis 8.3% Electives 10% Computer science 36.1%
c) Internet of things: The integration of the internet and the world of embedded systems is also one of the hottest topics these days. In this area, linking sensors to the internet is considered important. The area is emphasized by some German initiatives [4]. At Dortmund, special importance of this area results from the fact that TU Dortmund has logistics as a focus area and modern logistics is very much placing an emphasis on the internet of things. Introduction to logistics as well as other courses offered by the mechanical engineering department are on the list of electives. d) Robotics and control: Robotics and control for cyberphysical systems is a topic covered in many of the CPS research programs, such as NSF’s CPS program [5]. We include this area as a concentration. e) Systems of systems: In this area, we are considering the linking of of objects which by themselves are already systems. For example, we consider the linking of power stations to form a whole network of power stations and power consumers. Also, we consider networks of cars, trains or planes. Network technology is an essential ingredient of this area. Also, we must design means for locally safe operation despite possibly failing networks. At TU Dortmund, power systems would be a possible focus. Systems of systems are one of the focus areas of the European Commission. f) Life sciences: Our education in life sciences as a concentration focuses on the analysis of large amounts of medical data. Data analysis techniques form the kernel of this specialization. There would be some overlap with the specialization on big data, but there would also be some specific topics, on some cases linked to genetics and related topics in biology. 7) Global view: We plan to start with a list of electives taken from three focus areas and add more when the program becomes more mature. Note that all of the areas are based on some methodology which can be applied in various application domains. We could also image to have concentrations for certain application domains, like • transportation (automotive, avionics, rail), • power generation and distribution, • logistics, disaster planning and recovery, • automation, smart buildings, • bio-technology, health industries. Fig. 1 demonstrates the relative percentages of the involved disciplines.
Physics 19.4%
Fig. 1. Global percentages for undergraduate workload
C. Graduate courses There are many theoretical aspects of CPS. Our curriculum design is based on the idea of covering such aspects during the graduate education, both for computer science and physics. More depth and supplementary material is provided in CPSrelated topics. The material the theory of automata and parsing, as shown in Table VIII.
Name Topics in theoretical CS Elective: formal methods Elective: modeling & simulation Content computational game theory, complexity, automata, languages logic, formal methods for system design, functional programming discrete event systems, queuing Credits 8 9 9
TABLE VIII S UPPLEMENTARY CS MATERIAL DURING THE GRADUATE PHASE
We consider the emphasis on the analysis of physical experiments as one of our characteristics. hence, more depth is also included for physics (see Table IX).
Name Structure of matter Theoretical physics 1 Theoretical physics 2 Content solid state physics classical physics quantum mechanics Credits 9 9 9
TABLE IX S UPPLEMENTARY PHYSICS MATERIAL DURING THE GRADUATE PHASE
In addition, 24 credit points are reserved for electives. For example, advanced courses on data analysis, embedded system design or control may be selected. 30 credit points are allocated to the Master thesis, 6 credit points have to be earned in a project and 7 credit points are reserved for non-technical skills. A global break-down of the graduate phase is shown in Fig. 2 Note that many of the more theoretical courses are comprised in the graduate education. In this way, students are prepared to go for their PhD.
Non-technical skills 5.8% CPS project 5% Thesis 25% Physics supplements 22.5% Electives 20% CS supplements 21.6%
•
•
Fig. 2. Global percentages for graduate workload
III. R ELATED WORK Much of the experience with CPS education is based on embedded systems education. The workshops on embedded system education provide an overview over work in this area (see [6] for a recent edition). New educational CPS programs are being designed at several universities. At the time of writing, no complete overview of such programs is available. We expect a more complete view of the current situation to be one of the key results of the first workshop on CPS education. The book by Edward Lee [7] focuses on CPS education. IV. E VALUATION AND FEED - BACK Due to the fact that CPS is a brand-new area, no feed-back or evaluations by students exist for this area. However, the following evaluation and feed-back exists: • The call of the first workshop on CPS education describes a list of outcomes of a good CPS curriculum [8]. Due to the multidisciplinary nature of the proposed curriculum, we have addressed in particular the following suggested outcomes: A. “An ability to apply mathematical models of physical systems, cyber systems, and their composition”. D. “An ability to function effectively on multidisciplinary teams spanning cyber and physical domains.” E. “An ability to identify, formulate, and solve engineering problems that have both cyber and physical aspects.” G. “An ability to communicate effectively across cyber and physical domains.” J. “A knowledge of contemporary issues with cyber-physical systems.” K. “An ability to select and use appropriate techniques, skills, and modern engineering tools that span the cyber and physical domains.” Many of the other outcomes are also addressed, at least to some extent.
Our ES courses are designed to provide embedded systems foundations for cyber-physical systems. Therefore, experiences gained in this area can be applied. ES courses, even though they are electives, are very popular among students. Typically more than 100 students attend the German course and a smaller number the corresponding English course. No major difficulties have been reported. It is, however, essential that a course on computer organization is included in our program. Such a course is not present even in plans for a new curriculum for computer science education [9]. Also, most students attend a hardware lab. Prerequisites which might be lacking among CS students are included in the appendix of our ES text book [1]. In meetings where this program was presented, there was an unanimous agreement among colleagues to introduce this program. Colleagues wanted to introduce this program as soon a possible.
V. C ONCLUSION In this paper, we have presented our scheme for an integrated CPS curriculum. From the list of presented topics, it is clear that such an integrated program is required in order to allocate a sufficient workload. Such a program has to start already at the undergraduate level. With such an integrated program, there is headroom for teaching the fundamentals of the involved areas. However, even with such an integrated program it is not possible to include all the flavors of CPS. Instead of restricting ourselves to a particular flavor, we include these as areas for electives. ACKNOWLEDGMENT Part of the work on this paper has been supported by Deutsche Forschungsgemeinschaft (DFG) within the Collaborative Research Center SFB 876 “Providing Information by Resource-Constrained Analysis”. R EFERENCES
[1] P. Marwedel, Embedded System Design – Embedded Systems Foundations of Cyber-Physical Systems. Springer, 2011. [2] E. A. Lee, “Cyber physical systems: Design challenges,” in International Symposium on Object/Component/Service-Oriented RealTime Distributed Computing (ISORC), May 2008, invited Paper. [Online]. Available: http://chess.eecs.berkeley.edu/pubs/427.html [3] European Ministers of Education, “The bologna declaration of 19 june 1999,” http:// www.bologna-bergen2005.no/ Docs/ 00-Main doc/ 990719BOLOGNA DECLARATION.PDF, 1999. [4] acatech, “Cyber-physical systems — driving force for innovation in mobility, health, energy and production,” http:// www.fortiss.org/ fileadmin/ uploads/ projects/ agendaCPS Position-paper.pdf , 2011. [5] CPS Virtual Organization, “home page,” http:// www.cps-vo.org. [6] P. Marwedel, J. Jackson, and K. Ricks, Eds., WESE ’11: Proceedings of the 6th Workshop on Embedded Systems Education. New York, NY, USA: ACM, 2011. [7] E. A. Lee and S. A. Seshia, “Introduction to embedded systems, a cyber-physical systems approach,” http:// LeeSeshia.org, ISBN 978-0-55770857-4, 2011. [8] “First workshop on cyber-physical systems education — call for papers,” http://cps-vo.org/group/edu/workshop. [9] M. Sahami, D. Grossman, R. LeBlanc, and R. Seker, “Computer science curriculum 2013: Curricular guidelines for the next decade,” http:// cra.org/ uploads/ documents/ resources/ snowbird2012 slides/ sahami.pdf , 2013.